A method for the automatic classification of ECG beat on mobile phones

In this paper, we propose a novel method to perform the classification of electrocardiogram (ECG) beats on mobile phones. The Discrete Wavelet Transform and Higher Order Statistics are used to extract a set of eleven features from each ECG beat, and a Multilayer Perceptron is used to classify it between six types of beats. Our experiments show that a mobile phone can classify the ECG beats with an overall accuracy of 99.83%, and the sensitivity rates are greater than 99.48% for all beat. Running in a mobile phone phone, it needs only 27ms to classify each heart beat, what makes this method a reliable choice to perform computer-aided diagnosis on remote and critical situations.